On the uses of econometric models: A guide for policy makers

This paper reviews the evidence on ex ante and ex post forecasting with econometric models, considering in particular the role of human judgment in improving forecast accuracy. The importance of human judgment in econometric forecasting and the general superiority of ex ante over ex post forecasts provide evidence of model misspecification, which casts doubt on the validity of econometric policy simulations. An inadequate basis for selection among alternative models with differing policy implications undermines further the utility of these models in policy analysis and in testing hypotheses. Despite this record of performance macroeconometric models thrive and are widely used in policy analysis. Some reasons for their survival and importance in policy are suggested, and guidelines for the use of econometric models in the policy process are given.     

A User's Guide to State Revenue Forecasting

One of the most difficult tasks in preparing a state budget is forecasting available revenues. An overly optimistic forecast may subsequently require unpopular budget cuts, tax increases, or both. A pessimistic forecast may trigger a political controversy over the size of the budget surplus, or may encourage additional spending that cannot be sustained in the future. Forecasters have always informally shared their experience with each other. Only in recent years, however have scholars begun to document those experiences.
A forecast user is an elected official, program manager, reporter or private citizen involved in the budget process who uses a forecast prepared by someone else as a basis for making or influencing policy decisions about state taxes and spending levels. __ While the objectives of these forecast users may vary depending on their role in the budget process, they all share one need in common: a sound forecast which provides them with the information they need to participate in the budget process. The first part of this article briefly reviews the status of revenue forecasting in the states. The second part reviews five specific forecasting issues. The third section, discusses limits of revenue forecasting. The concluding section is a summary and checklist of good revenue forecasting practices.     

Econometric models and the study of the economic effects of Social Security

This article provides a relatively nontechnical discussion of previously published research on the use of econometric models in the study of the economic effects of social security. It illustrates the role of econometric model building by focusing on three major applications: forecasting, policy simulation, and hypothesis testing. A series of three macroeconomic examples serves to emphasize that the development and use of such models puts the focus of the analysis on the underlying economic structure. The first example presents a program-specific model of the Social Security system, the second a large-scale model of the U.S. economy, and the third a single-equation analysis of a specific issue.     

State Revenue Forecasts and Political Acceptance: The Value of Consensus Forecasting in the Budget Process

Concerns about political biases in state revenue forecasts, as well as insufficient evidence that complex forecasts outperform naive algorithms, have resulted in a nearly universal call for depoliticization of forecasting. This article discusses revenue forecasting in the broader context of the political budget process and highlights the importance of a forecast that is politically accepted—forecast accuracy is irrelevant if the budget process does not respect the forecast as a resource constraint. The authors provide a case illustration in Indiana by showing how the politicized process contributed to forecast acceptance in the state budget over several decades. They also present a counterfactual history of forecast errors that would have been produced by naive algorithms. In addition to showing that the Indiana process would have outperformed the naive approaches, the authors demonstrate that the path of naive forecast errors during recessions would be easily ignored by political actors.     

Voting and economic factors in French elections for the European Parliament

This article builds and estimates several econometric models that explain and forecast the outcomes of the French elections for the European Parliament. These models show the influence of the change in the local unemployment rate to explain the vote for the moderate Left and the moderate Right in the French elections for the European Parliament. These models appear to be accurate in forecasting the elections of the past, and their behavior for the 2009 French election for the European Parliament is satisfactory.     

Often Wrong,Never Uncertain: Lessons from 40 Years of State Revenue Forecasting

Experience in state revenue forecasting humbles and educates the public finance scholar and can inform the public administrator. It teaches the limits of econometrics, the importance of disaggregation, the significance of tax administrators, the utility of causal models, the issue of data problems, the need to understand tax structure, the importance of consensus forecasts, the terror of recessions, and the reality of being wrong. In the Indiana consensus system, experience provides greater respect for public servants seeking to make a sustainable fiscal system function and probably contributes to making the revenue forecast binding in the budget process.     

Beware of econometricians bearing estimates: Policy analysis in a “unit root” world

Current statistical approaches to modeling many economic relationships are grounded in traditional ideas of deterministic trends. Some of the failures of these approaches are due to inappropriate models using time series with “unit roots.” After a shock, unit root processes do not revert to some time trend, but rather can drift up or down without bounds. A “random walk” is a well-known example of a unit root process. The purpose of this paper is to explain the importance of unit root processes to policy analysts who make or rely upon econometric models using time series dat. In particular, the presence of unit root processes in GNP, energy and electricity consumption exports, imports, and other variables suggests that modifications to the way economic relationships are estimated may be necessary. Once these modifications are made, many important parameters turn out to be much different, with substantive implications for both forecasting and policy.     

Turning Point Analysis: Using Survey Research to Call Changes in Consumer Behavior and Predict Sales Taxes

Forecasters often disagree on revenue predictions because of differing techniques. Some states, such as Utah, have turned to the use of survey research methods to avoid this dilemma. This article reviews many of the national forecasting models examining consumer sentiment currently in use. And it compares the Utah experience with surveys used in other states.     

Municipal Government Revenue Forecasting: Issues of Method and Data

In this article the authors investigate the impact of the choice of time series method, the length of the data stream used to estimate the model, and the frequency of the data on forecasting accuracy for own source, non-tax general fund revenue for six Connecticut municipalities. The authors find that exponential smoothing models are generally the most accurate. They also conclude that local government officials should rely on bimonthly rather than monthly or quarterly data and retain, in a readily usable format, more than three years of data.     

Budgetary constraints and municipal revenue forecasting

Conclusion The objective of this research has been to determine whether key budget participants see forecasting as making significant contributions to the budgetary process. The issue is important for at least two reasons. Budgetary tools will be used to the degree budget players perceive them as providing net budgettary benefits. Moreover, the continuing trend toward technological sophistication gives governments opportunities, red herrings or not, to incorporate the technology into the budget process. The issue is also very timely; with the demise of packaged budget reforms but not the values of budget reforms, there may be new opportunities for improving budgeting on a less grandiose, more piecemeal basis, such as using forecasts to analyze budgetary options. The findings here provide some insight into two questions concerning governmental forecasting. First, why do governments use complex methods? In support of previous research, since cities most dependent on intergovernmental aid tend to use complex forecasting, such methods may indeed be seen as a way to help cope with fiscal stress. Also, reflecting the ambiguity of current research, since cities tend to use relatively simple techniques regardless of the revenue source being forecasted, the source is at best a partial determinant of complexity. The most important predictor of complexity, however, was budget format; cities that emphasize reform methods, especially planning, tend to use the most complex forecast methods. The forecasting process was not as important as expected.Second, so what? Does forecasting influence budgetary choices? The evidence from the second part of the study suggest that it can, but within definable limits. Budget directors are more likely than councils to value outyear estimates, but both actors are much less likely to value long-term estimates. The survey results also indicate that revenue forecasts are not as useful for making political decisions as for making management decisions: the forecast is usually used as an internal document, is only sometimes intended to affect council decisions, and is not usually included in the budget. In short, the forecast may be most useful for making managerial decisions since that is what most cities want out of it. It also tends to be more useful if the budget format is less traditional. Finally, the findings indicate that forecasting may be more useful to management to the degree the council finds it politically useful. This is extremely important since it suggests that as powerful as technology may be, budgetary tools that do not meet political needs will be managerially confined.Clearly, more research is needed in this area. Does forecasting actually shape long-range plans? In the long run, will the forecasting effort change the ways cities budget? Can the availability of forecasting information strengthen one actor relative to another? These are important questions that need answering to clarify the impact of forecasting and other technologies on the budget process.     

How Municipal Property Tax Responded to State Aid Cuts: The Case of Massachusetts Municipalities in the Post‐2001 Fiscal Crisis

This empirical study examines municipal property tax responses to state aid cuts during the post‐2001 fiscal crisis. Using panel econometric techniques on annual changes in property tax levy by 351 incorporated Massachusetts municipalities in the period of 2002–2006, the results suggest that municipal governments offset about 9 cents of each dollar of net state aid cut through the increase of their property tax rates. It contributes to public finance literature on the extent to which local property tax can be used to stabilize municipal revenues when the intergovernmental revenue shrinks.     

State Government Expenditure Determinants and Tax Revenue Determinants Revisited

The primary aim of the research was to test the general hypothesis that many institutional and political variables affect fiscal outcomes; that the preferences of public officials, and the decision processes used to act on those preferences, matter. State government data (state only, not state and local) from 1980, 1985, and 1990 were used to specify state tax revenue and expenditure models. A previous (Merrifield, 1991) state tax revenue model identified many significant institutional and political variables, but it was based only on 1985 data (49 observations). The larger data set (147 observations), including better data for some of the variables, indicated that most of Merrifield's (1991) findings are robust. The state expenditure models facilitated a second test of the general institutional and political variable hypothesis, as well as the hypotheses about specific variables. The expenditure model also facilitated comparisons of expenditure and tax determinants, and comparisons with results published in the literature. Though there are some noteworthy differences between the expenditure and tax models, the expenditure models also supported many of the institutional and political variable hypotheses.     

Monitoring Single-Source Revenue Funds throughout the Budgeting Process

Unexpected revenue shortfalls that occur during budget execution create serious problems for public agencies and programs. This is especially true for agencies and programs that rely upon a single source of revenue. This article implements a revenue monitoring process that combines ideas from statistical process control and exploratory data analysis into a spreadsheet environment. The methodology builds on principles used to monitor manufacturing processes. This simple application helps public managers detect revenue surpluses and deficits early enough in the budget execution process to take corrective action. The methodology also adapts to the data problems and avoids many of the pitfalls that arise when using "real world" budget data. An example of the benefits of this methodology is offered using an example of a state agency funded by a single revenue source.     

Presidential Elections and the Stock Market: Comparing Markov-Switching and Fractionally Integrated GARCH Models of Volatility

Bumba Mukherjee Department of Political Science, Florida State University, 554 Bellamy Building, Tallahassee, FL 32306 e-mail: smukherj{at}mailer.fsu.edu Existing research on electoral politics and financial marketspredicts that when investors expect left parties—Democrats(US), Labor (UK)—to win elections, market volatility increases.In addition, current econometric research on stock market volatilitysuggests that Markov-switching models provide more accuratevolatility forecasts and fit stock price volatility data betterthan linear or nonlinear GARCH (generalized autoregressive conditionalheteroskedasticity) models. Contrary to the existing literature,we argue here that when traders anticipate that the Democraticcandidate will win the presidential election, stock market volatilitydecreases. Using two data sets from the 2000 U.S. presidentialelection, we test our claim by estimating several GARCH, exponentialGARCH (EGARCH), fractionally integrated exponential GARCH (FIEGARCH),and Markov-switching models. We also conduct extensive forecastingtests—including RMSE and MAE statistics as well as realizedvolatility regressions—to evaluate these competing statisticalmodels. Results from forecasting tests show, in contrast toprevailing claims, that GARCH and EGARCH models provide substantiallymore accurate forecasts than the Markov-switching models. Estimatesfrom all the statistical models support our key prediction thatstock market volatility decreases when traders anticipate aDemocratic victory.     

Interstate Spillovers,Fiscal Decentralization,and Public Spending on Medicaid Home‐ and Community‐Based Services

This article examines the interstate spillover effect of Medicaid expenditures for home‐ and community‐based services (HCBS) and tests the relationship between fiscal decentralization and public spending. Based on the theory of interstate strategic interaction, an empirical model is specified that explicitly accounts for interdependence in states’ spending decisions. The model is estimated by applying spatial econometric methods to panel data for the 50 U.S. states for 2000–2010. Findings show a positive interdependence in state HCBS expenditures that is contingent on similarity in citizen ideology between states. Fiscal decentralization, measured by transfer dependence and revenue autonomy, is positively related to Medicaid HCBS spending.     

Fundamental models for forecasting elections at the state level

This paper develops new fundamental models for forecasting presidential, senatorial, and gubernatorial elections at the state level using fundamental data. Despite the fact that our models can be used to make forecasts of elections earlier than existing models and they do not use data from polls on voting intentions, our models have lower out-of-sample forecasting errors than existing models. Our models also provide early and accurate probabilities of victory. We obtain this accuracy by constructing new methods of incorporating various economic and political indicators into forecasting models. We also obtain new results about the relative importance of approval ratings, economic indicators, and midterm effects in the different types of races, how economic data can be most meaningfully incorporated in forecasting models, the effects of different types of candidate experience on election outcomes, and that second quarter data is as predictive of election outcomes as third quarter data.     

MANAGING MUNICIPAL SERVICES IN AN ERA OF DECUNING FEDERAL ASSISTANCE

Federal aid to local governments was cutback throughout the Reagan administration. In dealing with these revenue cutbacks, ex-recipients have a variety of options, to find new sources of revenue or reduce expenditures. Based on the perspectives of local government budget officials and environmental data, this study attempts to determine which governments will be most successful in minimizing service cutbacks in this environment. Using a cross-sectional analysis and data concerning federal revenue sharing, we find both support and differences for many predictors found in previous case-studies of fiscal stress. The differences likely stem from the new cutback environment. This study illustrates that management strategies under fiscal decentralization are strategic and not haphazard.     

A critique of quasiexperimental and statistical controils for measuring program effects: Application to urban growth control

Most studies evaluating the effect on housing prices of regulatory programs for controlling urban growth use econometric methods (hedonic price models) without considering the research-design aspects of the task. The article examines the strengths and weaknesses of several quasiexperimental and statistical methods for measuring the effect of growth-control programs, using criteria extracted from theory. The methodological critique is tested empirically by comparing several methods using data from the growth-control community of Davis, California and from three comparison communities. The empirical test demonstrates that different research designs can produce, large differences in the magnitudes of estimated effects, and that adding statistical controls to quasiexperimental controls permits the detection of smaller effects.     

Price Setting for Local Government Service Delivery : An Exploration of Key Issues

Local governments in Australia are constantly seeking to raise additional revenue to fund higher service demands. One key revenue source is service fees and charges, including fines. Premised on the notion of user pays, service fees and charges represent a significant proportion of total revenue for many local governments, especially in New South Wales, Queensland and Tasmania. This exploratory essay addresses a number of key issues related to this revenue source in order to stimulate discussion and debate on matters which are presently under‐examined in the literature. The article examines the philosophy underpinning price setting, the identification of the principles of price setting, the adoption of applicable price setting models, the need to adhere to National Competition Policy and also the applicability of differential pricing of service delivery within local government. Calls for more open approaches and enhanced disclosure relating to service provision and pricing are made.     

The British general election: Synthetic forecasts

Election forecasting work, in Britain and elsewhere, has been confined mainly to traditional approaches – statistical modeling or poll-watching. We import a new approach, which we call synthetic modeling. These models, developed out of forecasting efforts from the American 2012 presidential election, we are currently testing in a comparative European context. In this exercise, we apply the strategy to forecasting British general elections, in particular the upcoming 2015 contest. These synthetic forecasts come from hybrid models blending structural knowledge with contemporary public opinion. We generate ongoing nowcasts, from six months prior, to one month before, Election Day itself.